For over half a century, androgen deprivation therapy (ADT) has been known to have beneficial effects on prostate cancer due to the testosterone dependence of prostate tumor cells, yet its optimal use still remains undefined. Historically limited to palliative treatment, the 1990s witnessed a dramatic growth in its use across a variety of clinical scenarios, culminating in nearly half of men with prostate cancer receiving ADT within a year of diagnosis and over a quarter of prostate cancer survivors being on androgen deprivation in any given year. More recent developments have changed the landscape for ADT further and pose an ongoing challenge to optimizing its use for prostate cancer. Clinical trials published in 2002 and 2004 have clearly established an overall survival benefit for the use of ADT as an adjuvant to radiation in men with locally advanced or high risk localized disease. However, even as evidence of efficacy for ADT has become clearer, there has been increasing awareness since 2005 of potentially life-threatening adverse effects such as osteoporosis, bone fractures, cardiovascular disease and development of diabetes. One challenge is therefore to determine which patients are more likely to benefit rather than be harmed from use of ADT. A second issue is that these newly noted complications of ADT fall outside the usual purview of cancer care. The presence of a primary care provider (PCP) or a model where patients see both a POP and cancer specialist may therefore help mitigate the adverse effects of ADT. To address these issues, we propose to take a population-based approach, using SEER-Medicare data, to achieve the following specific aims: 1) identify subgroups of patients at particular risk of complications from ADT and 2) Examine the effect of model of care delivery (cancer specialist dominant vs. POP dominant) on care of men on ADT.